Troubleshoot OpenClaw Model Failover and Auth Errors

We have all been there: you are running a complex task and suddenly everything stops because an API provider hit a rate limit or had a brief outage. It is frustrating to have your workflow interrupted by something outside your control. You want your system to stay up and keep working without manual intervention.

OpenClaw handles these hiccups through a failover system, but sometimes the configuration needs a little attention to work correctly.

How does failover work?

Failover happens in two stages:

1. Auth profile rotation within the same provider
2. Model fallback to the next model in agents.defaults.model.fallbacks

Cooldowns apply to failing profiles using exponential backoff. This ensures OpenClaw can keep responding even when a provider is rate-limited or temporarily failing.

What does “No credentials found for profile ‘anthropic:default’” mean?

It means the system attempted to use the auth profile ID anthropic:default, but could not find credentials for it in the expected auth store.

Fix checklist for “No credentials found for profile anthropic:default”

1. Confirm where auth profiles live:

   • Current: ~/.openclaw/agents/<agentId>/agent/auth-profiles.json
   • Legacy: ~/.openclaw/agent/* (migrated by openclaw doctor)

2. Confirm your env var is loaded by the Gateway

   • If you set ANTHROPIC_API_KEY in your shell but run the Gateway via systemd/launchd, it may not inherit it
   • Put it in ~/.openclaw/.env or enable env.shellEnv

3. Make sure you’re editing the correct agent

   • Multi-agent setups mean there can be multiple auth-profiles.json files

4. Sanity-check model/auth status

   openclaw models status

Fix checklist for “No credentials found for profile anthropic”

This means the run is pinned to an Anthropic auth profile, but the Gateway can’t find it.

Use a setup-token:

claude setup-token
openclaw models auth setup-token --provider anthropic

If you want to use an API key instead:

• Put ANTHROPIC_API_KEY in ~/.openclaw/.env on the gateway host
• Clear any pinned order:

  openclaw models auth order clear --provider anthropic

Confirm you’re running commands on the gateway host

• In remote mode, auth profiles live on the gateway machine, not your laptop

Why did it also try Google Gemini and fail?

If your model config includes Google Gemini as a fallback (or you switched to a Gemini shorthand), OpenClaw will try it during model fallback. If you haven’t configured Google credentials, you’ll see No API key found for provider "google".

Fix: either provide Google auth, or remove/avoid Google models in agents.defaults.model.fallbacks.

If you are still having trouble getting your keys recognized, check out the AI Setup Assistant.

Next steps

• Model Configuration Guide
• Gateway Setup